CentOS - May 2011 - EL 6 rollout strategies? (Scientific Linux)

I was wondering what feedback might be offered by the CentOS community on their 
experiences using Scientific Linux? 

I'm a long-time Centos user, and am basically happy with CentOS. I
understand
there are delays getting EL 6 out.  We have been long anxious to roll out EL 6 
as soon as it's ready, but our time window for rollout is looming and we
will
need to act. (for business reasons, we need to rollout over summer, before Aug 
1, we need to start regression testing now!) 

I was once a WhiteBox Enterprise Linux user and switched to CentOS 4 without 
issue, and am assuming that I might need to do something similar if we decide 
to go this route. 

Any feedback is appreciated! 

-- 
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.

-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.centos.org/pipermail/centos/attachments/20110506/85018191/attachment-0005.html>

centos-bounces at centos.org wrote:
> I was wondering what feedback might be offered by the CentOS
> community on their experiences using Scientific Linux?
Fresh install of 6.0 without a hitch a while ago.


Insert spiffy .sig here:
Life is complex: it has both real and imaginary parts.

//me
*******************************************************************
This email and any files transmitted with it are confidential and
intended solely for the use of the individual or entity to whom
they are addressed. If you have received this email in error please
notify the system manager. This footnote also confirms that this
email message has been swept for the presence of computer viruses.
www.Hubbell.com - Hubbell Incorporated**

On 05/06/2011 01:31 PM, Benjamin Smith wrote:
> I was wondering what feedback might be offered by the CentOS community
> on their experiences using Scientific Linux?
> 
> 
> I'm a long-time Centos user, and am basically happy with CentOS. I
> understand there are delays getting EL 6 out. We have been long anxious
> to roll out EL 6 as soon as it's ready, but our time window for rollout
> is looming and we will need to act. (for business reasons, we need to
> rollout over summer, before Aug 1, we need to start regression testing
> now!)
> 
> 
> I was once a WhiteBox Enterprise Linux user and switched to CentOS 4
> without issue, and am assuming that I might need to do something similar
> if we decide to go this route.
> 
> 
> Any feedback is appreciated!
We are getting fairly close to having a tree ready to send to QA.  The
goal for sending the tree is 10 May 2011.  It might not happen before
then, but it should happen within a week of that date.

Disclaimer:  We may have something that fails to work and throws a
monkey wrench in the plans ... but it is getting close.

I would expect once it is in QA that we can release in 2-4 weeks
(maximum) from that point.

But the real question is, do you want to use EL6.  I personally would
only roll out testing stuff on EL 6 at this point (be it SL 6.0, Oracle
UBL 6.0, RHEL 6.0, etc.).  CentOS 5 still has 3 years of normal support
before its retirement date, and is much more mature at this point (IMHO).

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 253 bytes
Desc: OpenPGP digital signature
URL:
<http://lists.centos.org/pipermail/centos/attachments/20110506/d1dd6a05/attachment-0005.sig>

On Fri, May 6, 2011 at 19:31, Benjamin Smith <lists at
benjamindsmith.com>wrote:
> I'm a long-time Centos user, and am basically happy with CentOS. I
> understand there are delays getting EL 6 out. We have been long anxious to
> roll out EL 6 as soon as it's ready, but our time window for rollout is
> looming and we will need to act. (for business reasons, we need to rollout
> over summer, before Aug 1, we need to start regression testing now!)
>
>
So buy RHEL.

-- 
Kind Regards,
Christopher J. Buckley
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.centos.org/pipermail/centos/attachments/20110506/4b27479a/attachment-0005.html>

> I'm a long-time Centos user, and am basically happy with CentOS. I
> understand there are delays getting EL 6 out. We have been long anxious to
> roll out EL 6 as soon as it's ready, but our time window for rollout is
> looming and we will need to act. (for business reasons, we need to rollout
> over summer, before Aug 1, we need to start regression testing now!)
Hi !

Here, we have a cluster of 8 nodes that we just deployed in RHEL 6 (the real 
thing, 14 k$/year). And with that, came other servers (router, test servers, 
developpement servers), and I wanted to have also the same OS. While waiting 
for C6, I installed an unsubscribed version of RHEL6, but it was troublesome 
to install packeges. So for those servers that were already installed, I 
switched them to SL6 without having to re-install, and it went great without 
a pain. I installed some other servers with SL6, and didn't noticed any 
difference.

I plan to switch those SL6 servers back to C6 when it's out, for uniformity 
reasons, I don't anticipate any problems then.

Regards,

On 5/6/2011 1:44 PM, Johnny Hughes wrote:
>
> But the real question is, do you want to use EL6.  I personally would
> only roll out testing stuff on EL 6 at this point (be it SL 6.0, Oracle
> UBL 6.0, RHEL 6.0, etc.).  CentOS 5 still has 3 years of normal support
> before its retirement date, and is much more mature at this point (IMHO).
In this business, mature is a synonym for ancient.  But the bigger issue 
is that 3 years out may not cover the lifespan of a newly deployed 
server that you'd really like to keep running with nothing more 
complicated than 'yum update'.

-- 
   Les Mikesell
    lesmikesell at gmail.com

On 05/07/2011 09:00 AM, Benjamin Smith wrote:
>
> I was wondering what feedback might be offered by the CentOS community on
their
> experiences using Scientific Linux?
>
> I'm a long-time Centos user, and am basically happy with CentOS. I
understand
> there are delays getting EL 6 out.  We have been long anxious to roll out
EL 6
> as soon as it's ready, but our time window for rollout is looming and
we will
> need to act. (for business reasons, we need to rollout over summer, before
Aug
> 1, we need to start regression testing now!)
>
> I was once a WhiteBox Enterprise Linux user and switched to CentOS 4
without
> issue, and am assuming that I might need to do something similar if we
decide
> to go this route.
>
> Any feedback is appreciated!
I had to complete a new VM Host machine before CentOS-6 was ready, so I 
used SL-6 as the underpinnings.  It simply worked, once I ironed out a 
couple of unrelated hardware issues.  It auto-updates and lets me know 
when that's been done, but it doesn't reboot unless I want to, so 
there's little risk in allowing the updates.  Auto-update can be 
disabled if you prefer.

So far it has been flawless.  It's managing two very large software 
RAID-6 arrays and 7 Guest VMs on a dual-Xeon Supermicro motherboard.

All of the VM Guest OS's are CentOS-5.6 which works well for my 
applications.  I may or may not upgrade the Guests to C6, depending on 
need.  CentOS has been so reliable that I'm reluctant to move to a 
different platform.  I've used Ubuntu Server and it works well, but I'm 
more familiar with the RHEL way of doing things so I'll stick with CentOS.

One thing I don't know much about is an in-place upgrade of C5 to C6. 
There are some under-the-hood differences that must be taken into 
account, which I haven't looked into.

YMMV,
Chuck

On Tuesday, May 10, 2011 09:17:39 PM Craig White wrote:
>  Upstream released exactly 6 months ago and still
> nothing and apparently today's target date has slipped, and 2) until
> CentOS admits that there is a problem, nothing will actually change.
Please read the CentOS-devel list and IRC channel.  There are some changes going
on WRT visibility of the process, and time will tell if that sticks.

My gut feel, not being one of the developers doing this, is that once the
package build order and buildroots are figured out for 6.0 that 6.1 should be
far less work.  But I reserve the right to be wrong.

How long it will take is of course anyone's guess; after all, it's been
quite a while since 5.6's release, and SL, as fast as they were with 6.0,
doesn't have a 5.6 full release out (beta 2 is due this Friday, but
that's a beta and not a production release.  Of course, they've also
backported security fixes where possible from 5.6 back to 5.5, but that's
part of their policy, plan, and procedures).

To get these things right takes time.  CentOS spent the time up front on 5.6 and
4.9, and both of those were released non-beta before SL released those versions;
SL has since released 4.9.  Both projects are doing a fantastic job of trying to
nail the proverbial blob of gelatin to the wall, and I've hesitated
comparing them in any way, simply because I don't want to disparage either
project.  And the two projects are not in competition.  And neither project has
a fully visible buildsystem.

In my case, I have essentially three choices:
1.) Use SL 6;
2.) Wait on C6;
3.) Buy RHEL6.

All of the three have costs, visible and hidden.  3 obviously has monetary
costs, but both 1 and 2 have time and risk costs, since neither SL nor CentOS
will be as fast on updates as choice 3.

There are boxes I'm possibly going with SL, but my servers are likely to
remain CentOS, unless and until I can get funding to purchase RHEL (which, since
it's a subscription, must be purchased out of opex funding).  But I fully
realize that if I want a fully supported product in the EL space I'm going
to have to pay for it, either with RHEL or Oracle or SuSE.  Otherwise I'm
going to be happy with what I get, even if that's late.

On Wednesday, May 11, 2011 01:51:08 PM Les Mikesell
wrote:
> I've always been a fan of the 
> coordination they have among the additional repositories that is lacking 
> in yum/rpm equivalents and was impressed when my 9.0.4 installs 
> painlessly upgraded themselves to 10.0.4.  
You must not have many PPA's enabled.  And you must not use PostgreSQL,
which won't painlessly upgrade on anything.....
> Admittedly, not as many 
> locally configured apps as on my Centos boxes, but it all still seemed 
> to be working after the major-version over-the-network upgrade.
I've had the opposite experience with several clients, using Ubuntu as a
desktop, not a server.  I've had a few issues with servers, too.

Timely updating takes effort; either I pay with money for upstream's
binaries or I pay with time for either upstream's source RPMs (which can be
delayed) or a rebuild's binaries.  Or I pay with transition cost to a
different distribution.  Those are the choices. TANSTAAFL.

[drifting farther off-topic....]

On Wednesday, May 11, 2011 04:34:49 PM Les Mikesell
wrote:
> On 5/11/2011 3:18 PM, Lamar Owen wrote:
> >  And you must not use PostgreSQL, which won't painlessly upgrade
on anything.....
> 
> Automatically doing the dump/load (and magically finding the space for 
> it) for version changes that need it would be a lot to ask.
Yes, I know.  Tried.
> >  I've had a few issues with [Ubuntu] servers, too.
> 
> With the LTS versions?  
Yes.

One upgrade I did from C4 to C5 (with upgradeany) was smoother than the last LTS
upgrade I tried.  I liken the C5 -> C6 upgrade path as trying to take a
Ubuntu LTS 6.06 to a 10.04; which path I tried, and failed, to get working.  In
one case it was with a Dell laptop that came with Ubuntu from Dell, and that is
supported by Dell with Ubuntu.  Sound quit (known issue), wireless went funky. 
One 'accidental' (client-initiated) upgrade from 8.04 to 10.04 lost
keyboard and mouse after gdm got control.

And even with Dell's that have RHEL support, I've seen issues with
CentOS upgrades; but, then again, neither CentOS nor RHEL ( nor SL) support
upgrading.

Upgrades are difficult problems to solve, and at the moment I don't know of
any distribution (that claims upgradability) that gets it completely right for
all the cases I've tried.

The CentOS path (it's not supported, but if you're brave and know
exactly what you're doing there is upgradeany to let you shoot yourself in
the foot) I feel is the correct one.

On Thursday, May 12, 2011 06:23:52 AM Christopher Chan
wrote:
> 6.04->10.04? Nah, you are supposed to jump to 8.04 and then to 10.04.
I did 6.06 -> 8.04 -> 10.04, and it broke.  Badly.
> > Upgrades are difficult problems to solve, and at the moment I
don't know of any distribution (that claims upgradability) that gets it
completely right for all the cases I've tried.
> 
> Not even Debian?
The one box I ran Debian on was somewhat unusual, and I lost SMP capability on
the box upgrading the one time I did.  The box is a DEC AlphaServer 2100
(Sable), and Sable SMP is hard to get these days; the last Debian kernel I know
of that supported it was a 2.2 series kernel......

Looking more like I'm going to do at least one, and possibly more, SPARCs on
Debian, I'll get a little more experience with it then.  I'd rather do
CentOS, and be consistent across servers in terms of administration....
> On the OpenSolaris/OpenIndiana side of things, I have not had problems. 
> And you get a complete rollback option too as a bonus.
Rollbacks would be good.

The smoothest upgrades of any I've ever done have been on EMC Clariion gear
and its FLARE operating environment.  But that's controlled hardware, and
tightly controlled software, not a general purpose OS, so not directly
comparable.

On Tuesday, May 10, 2011 09:12:59 PM Dag Wieers wrote:
> And given that C5.6 took 3 months, are there any reasons why C6.1 would 
> take no more than 1 month ?
I can easily think of a few.  4.9 and 6.0 are two of those few.

Again, I'll note that SL is just now releasing the second beta of 5.6 this
Friday.  Not the final SL 5.6.  Not to try to make SL look bad in any way, but
to highlight that the triple-threat of 6.0, 5.6, and 4.9 is/was hard on both
projects.

CentOS chose to do 5.6 and 4.9 before 6.0; CentOS is later with 6.0.  SL chose
to do 6.0 first; they had 6.0 out first, and 4.9 followed, and 5.6 is not yet
out.  I don't consider a beta or alpha or even an RC release as being
'out' either.

And again I'm not being critical of either project; both do a fantastic job.
They had and have different priorities, and with a CentOS 5.6 release behind us
we now see the effect of those different priorities.

Who cares? I find it amazing that these guys still keep on building and
providing considering how their "users" treat them.

Team CentOS, keep your heads up. For me, you are still the best thing
that happened since sliced bread.

Come on, community, where is your love?

My 2 pence,

Janne "Janski" AKA JNixus Nyman
Founder of Newman IT Solutions Ltd


-----Original Message-----
From: centos-request at centos.org
Reply-to: centos at centos.org
To: centos at centos.org
Subject: CentOS Digest, Vol 76, Issue 16
Date: Mon, 16 May 2011 12:00:02 -0400

Re: [CentOS] EL 6 rollout strategies? (Scientific Linux)

On 05/16/11 2:41 PM, Radu Gheorghiu wrote:
> I never said I want to do it.
ah, so what DID you say?  you want someone unspecified to do a 
better/different job for you than someone else is already doing for free ?

man, its easy to volunteer other people from the comfort of your desk.

-- 
john r pierce                            N 37, W 123
santa cruz ca                         mid-left coast

On Tuesday, May 17, 2011 07:56:59 AM Jerry Franz wrote:
> On 05/17/2011 03:06 AM, John Doe wrote:
> > Maybe all the non-technical discussions could go into a "CentOS
> > Politics/Philosophy" new list...?
> 
> And on that note, some required reading for everyone in this floating 
> flame war. Don't skim it - read it.
That is a good read; particularly the piece about the probability of a group
asking for a moderator is directly proportional to the age of the group (not an
exact quote).

Thanks for posting.

On Wednesday, May 18, 2011 09:23:14 AM Brunner, Brian T.
wrote:
> Rebuilding somebody else's sources without their build environment
isn't
> typical.  It's MindReading 101.
It's worse than that in the specific case of EL6.  It's replicating the
result without replicating the build system.  It's a pretty well-known thing
that upstream is building with Koji fed from a source code management system; 
CentOS is not as far as we know (and it's overkill anyway, unless you add
several things to the distribution as SL does, and they're using Koji for
SL6, and started learning Koji and setting up their buildsystem for 6 nearly a
year ago).  Koji in fact will not allow, by default, a 'normal' user to
rebuild from source RPM, but requires building out of the SCM for normal users. 
The case of a 'from source RPM' rebuild is not Koji's forte.

It's also fairly well-known that mock builds in koji and mock builds outside
of koji can sometimes differ.  Grep the archives of several lists to verify
that; I've seen it before, but I don't have time at the moment to pull
up the reference.  I have a VAX to redisk and boot up....